Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cleanup LSP error reporting #74530

Merged
merged 1 commit into from
Aug 14, 2024
Merged

Conversation

dibarbet
Copy link
Member

@dibarbet dibarbet commented Jul 23, 2024

Fixes https://devdiv.visualstudio.com/DevDiv/_workitems/edit/1922370

We were reporting connection lost exceptions when the server shutdown due to refresh notifications being sent during shutdown (they are triggered by events that are not a part of the shutdown handling in the queue).

This modifies the code to catch connection lost exceptions from refresh notifications. Other parts of the system already handle connection loss issues and report failures.

Additionally fixes an issue where we would not report NFW for mutating LSP requests.

@dibarbet dibarbet requested a review from a team as a code owner July 23, 2024 23:51
@dotnet-issue-labeler dotnet-issue-labeler bot added Area-IDE untriaged Issues and PRs which have not yet been triaged by a lead labels Jul 23, 2024
{
if (rethrowExceptions)
{
return nonMutatingRequestTask;
try
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

fixes a bug here where we would only report NFW for non-mutating requests. If a mutating request failed we would never get fault telemetry for it.

await nonMutatingRequestTask.ConfigureAwait(false);
}
// If we had an exception, we want to record a NFW for it AND propogate it out to the queue so it can be handled appropriately.
catch (Exception ex) when (FatalError.ReportAndPropagateUnlessCanceled(ex, ErrorSeverity.Critical))
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i believe this is the correct pattern to report and throw, but let me know if not.

});

var didReport = false;
FatalError.OverwriteHandler((exception, severity, dumps) =>
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tests to verify NFW handler is called for mutating and non-mutating requests

@dibarbet dibarbet merged commit 74270d4 into dotnet:main Aug 14, 2024
25 checks passed
@dibarbet dibarbet deleted the fix_shutdown_error branch August 14, 2024 20:58
@dotnet-policy-service dotnet-policy-service bot added this to the Next milestone Aug 14, 2024
@dibarbet dibarbet modified the milestones: Next, 17.12 P2 Aug 26, 2024
dibarbet added a commit that referenced this pull request Nov 15, 2024
…ve async listener in queue (#75907)

Should resolve
https://dev.azure.com/devdiv/DevDiv/_workitems/edit/2224584

Optprof web application tests have been failing for a while in VS. Tim
tracked down the issue to an async listener introduced in
#74530

The never completing listener were requests for workspace diagnostics.
This is actually expected for workspace diagnostics requests, as they
are held open indefinitely if nothing changes (see
https://github.com/dotnet/roslyn/blob/main/src/LanguageServer/Protocol/Handler/Diagnostics/AbstractWorkspacePullDiagnosticsHandler.cs#L94).
So as long as workspace diagnostics requests function this way, it is
easy to get into a situation where there are async listeners that never
complete.

The queue async listener was introduced to ensure that unit tests could
reliably verify if the NFW handler was called (as handling the exception
happened after the response was returned to the client).

To resolve this issue I did a couple things
1. Move reporting NFW for request exceptions into the telemetry
reporting scope. Importantly this scope is called before the result is
sent back to the client. This allows unit tests to simply wait for the
request to complete before checking if the NFW handler was called (no
need for an async listener).
2. Removing the async listener from the queue now that no unit tests
need to wait for an action to occur in the queue.


TODO - validate optprof
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Area-IDE untriaged Issues and PRs which have not yet been triaged by a lead
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants